Prayer and Healing at St. Luke's.

As almost any poll will show, an over-whelming number of Americans believe in the power of prayer--specifically, in prayer's alleged effectiveness in healing disease. By the same token, many world cultures attribute sickness to sin, making prayer or some related ritual an accepted healing modality among the people. This power, however, isn't said to inhere only in prayers uttered by the sick on their own behalf On which case a salutary result may lend itself to a scientific explanation, such as the "placebo effect" or autosuggestion). Intercessory prayers--prayers made for the sick by others, including prayers of which a sick person is completely unaware--are also said to be effective.

But scientific evidence for the efficacy of intercessory prayer has been conspicuously lacking. The one published study frequently cited on the alleged positive effects of intercessory prayer is that of Dr. Raymond C. Byrd: "Positive therapeutic effects of intercessory prayer in a coronary care unit population," published in 1988 in the Southern Medical Journal. Byrd tantalizingly reports that intercessory prayer had a beneficial effect on the course of heart disease in 393 patients in a San Francisco coronary care unit. I say "tantalizingly" with good reason, because that research hasn't been successfully replicated, despite frequent calls from both supporters and detractors of the study for such repetition.

At least, not until now.

On the morning of October 26, 1999, the world awoke to an Associated Press headline, "Study: Prayer Helps Heart Patients" (no, modifiers such as claims or may), over the byline of Brenda C. Coleman. The day before, the Archives of Internal Medicine, a publication of the American Medical Association, had published a report by eight authors of an attempt to replicate Byrd's findings. Entitled "A randomized, controlled trial of the effects of remote, intercessory prayer on outcomes in patients admitted to the coronary care unit," it was a one-year study on nearly 1,000 patients at the Mid America Heart Institute (MAHI) of St. Luke's Hospital in Kansas City, Missouri. The patients were divided into two groups: the experimental group which was prayed for, and the comparison or control group which was not. The gist of Coleman's news story is that "heart patients who had someone praying over them without their knowledge suffered 10 percent fewer complications." The lead author of the scientific report, William S. Harris, is quoted as saying, "It's potentially a natural explanation we don't understand yet. It's potentially a super - or other-than-natural mechanism." In lay language: maybe it isn't a miracle, maybe it is.

If it is a miracle, it constitutes a poor showing by the Almighty, bestowing as he does a measly 10 percent improvement upon prayed-for patients. And Harris concedes that his study had its limitations. As Coleman notes, "Among other things, many patients in the comparison group likely had friends and relatives praying for them, too." Coleman doesn't seem, however, to have considered the full ramifications of what Harris had admitted--namely, the possibility that any result the study might have obtained could have been explained merely by invoking various numbers of unknown intercessors, to the ends of the Earth, praying for persons in one group or the other, or for people in general, for human happiness, or whatever.

Another criticism, potentially more damaging, was voiced by Harvard University's Herbert Benson, president of the Mind/Body Medical Institute at Beth Israel Deaconess Medical Center in Boston, Massachusetts. Benson cautioned that in both this and the earlier Byrd study, the results had been scored by new systems, different from each other and neither as yet medically validated. And if these systems cannot ultimately be validated, the results of the studies using them will be meaningless.

Looking at the Archives of Internal Medicine article, I can see that in this study, on the surface, all the usual precautions were taken. The study's objective was "to determine whether remote, intercessory prayer for hospitalized, cardiac patients will reduce overall adverse events and length of stay." How this was done is described, one phrase at a time, in operational terms.

The design was that of a "randomized, controlled, double-blind, prospective, parallel-group trial." Patients were admitted consecutively to the coronary care unit (CCU) and randomized into a "prayer group" (the even-numbered medical records) and a "usual care group" (the odd-numbered medical records). Those readmitted kept their previous record numbers; those admitted for workup or for cardiac transplantation were excluded, as were those released within twenty-four hours. Of the remainder--990 in all--466 got in the prayer group while 524 ended up in the usual care group. (The probability that this size difference would occur by chance is 18 percent, well within the conventional limits for chance occurrences.) Informed consent wasn't sought and "neither patients nor medical staff [were] aware that a study was being conducted."

The seventy-five intercessors, fifteen teams of five each, were Christians recruited from the community. The largest group of these--representing 35 percent--was "nondenominational" (a term often used to denote evangelicals); the next largest--27 percent--was Episcopalian. All agreed with the statements: "I believe in God. I believe that He is personal and is concerned with individual lives. I further believe that He is responsive to prayers for healing made on behalf of the sick." Each team of five intercessors had a leader, who was given only the first names of new prayer-group patients to share with the other four team members. All five were asked to "pray daily for the next 28 days [the maximum hospitalization time of 95 percent of the patients] for `a speedy recovery with no complications' and anything else that seemed appropriate to them." Patients' requests for prayer from the chaplain's staff--5 percent--were always honored.

Up to this point, except for the patient randomization procedure, which I will discuss below, the experimental design wasn't particularly remarkable. The evaluation of "clinical outcomes," however, is another story. For if you pray to receive a bicycle or a Mercedes-Benz within a specified time, there's usually no problem determining whether or not the prayer has been answered. In the case of this study, however, there was a built-in difficulty. According to the article:

   Since prayer was offered for a speedy recovery with no complications, it
   was anticipated that the effect of prayer was unlikely to be evident in any
   specific clinical outcome category (e.g., the need for antibiotics, the
   development of pneumonia, or the extension of infarction), but would only
   be seen in some type of global score. Review of the medical literature
   revealed no previously validated and standardized statistic to quantitate
   severity of outcomes in critically ill cardiovascular patients.

Therefore, as in the 1988 Byrd study, the investigators developed their own system of evaluation: "a weighted and summed scoring system called the MAHI-CCU score."

This kind of de novo creation of a new research instrument is always risky, and it isn't clear to me why the authors couldn't have modified their specific aims so they could employ methods previously validated (such as those for testing specific medicines, procedures, and so forth). The list of "diagnoses, events, and procedures" used for evaluating the effects of prayer was an omnium-gatherum of things that might happen or be done to any patient in the CCU. These were further weighted according to severity on a scale of one (unstable angina) to six (death). It also isn't clear to me how it would have been possible to avoid counting some events, at least indirectly, more than once, but I must leave that judgment to cardiologists. I anticipate that future editions of the Archives of Internal Medicine will include some--I hope many--letters from readers with extensive clinical experience in this field, which should help to shed further light on this aspect of the methodology.

But one doesn't have to be a cardiologist or rocket scientist to know that, no matter how honest and upright the "blinded" participants in this research may have been, it ordinarily would be very difficult to maintain total ignorance of the relevant circumstances for an entire year among such a large group of people. Furthermore, one especially odd feature of the experimental design was having the chaplain's secretary be responsible for the assignment of patients to the two groups--according to whether their chart numbers were odd or even. Certainly a computer could have done at least as well and with more assurance of objectivity. These are just two examples of what seems to me a general disregard or denial of obvious security issues in research of this kind.

George Robert Price addressed this problem in a notorious 1955 paper, "Science and the Supernatural," published in Science. Emboldened by the words of David Hume that "no testimony is sufficient to establish a miracle, unless the testimony be of such a kind that its falseness would be more miraculous than the fact which it endeavors to establish," Price called on the scientific community to demand only the most rigorous standards of evidence from all investigators claiming to have observed phenomena--in this case those collectively known as "extrasensory perception"--that appear to violate known laws of physics. He went on to suggest--unwisely, in the opinion of many who otherwise agreed with him--that the possibility of deliberate fraud must be considered if all other explanations fail.

At the time I was among those shocked by my colleague's unmannerly attitude, which seemed so egregious that Time picked up his argument ("Challenge to Psi") in its next issue. Price was subsequently lionized for his bravery (or foolhardiness) by some in the scientific community and shunned by others--perhaps the majority--for his indiscretion. Many years later, however, I was forced to change my reaction. It was after a stint as acting associate chief of staff for research at a Veterans Administration medical center and later as chair of its research committee, and especially after making the painful discovery that a worker in my own laboratory had knowingly submitted "adjusted" data to me, that I realized that it can be a good deal harder to detect folly or fraud than most scientists would like to think.

The power and beauty of careful scientific inquiry lies in its self-corrective nature, which we can never take for granted. It is especially important to be critical of the design of certain kinds of experiments that by their nature (complexity, cost, unprofitability to potential investigators, and so on) are unlikely ever to be replicated sufficiently to be truly self-correcting. In such cases, Hume's dictum that "extraordinary claims require extraordinary evidence" might well be clarified by specifying "obtained under extraordinary security" as a part of the requirement.

Despite these issues, however--and despite the fact that Harris also publicly defends the notion of "intelligent design" against naturalistic evolution--it seems to me that the specific findings reported in the Harris et al. prayer study, as well as the discussion of possible interpretations and the conclusions reached, were all appropriately modest. The authors were careful to note that they "have not proven than God answers prayer or that God even exists." Additionally, it is of no little interest that, although the stated purpose was to "attempt to replicate Byrd's findings," the conclusion is somewhat contradictory: "Our findings support Byrd's conclusions despite the fact that we could not document an effect of prayer using his scoring method."

Such a study as this can, I believe, serve a useful purpose if it leads to more studies--more carefully designed and executed--and to greater critical attention to the questions it attempts to raise. According to the Knight Ridder report of October 31, 1999, bylined by Eric Adler, "In Boston, Harvard University researcher Herbert Benson is conducting a study of the effects of intercessory prayer on 1,800 patients admitted for heart surgery." The results of that study are expected late in 2000. So I hope to be able to follow this type of research to something medical scientists can look upon as a successful conclusion.

David Schafer holds a Ph.D. in physiology from the University of Minnesota and is .a member of the American Humanist Association board of directors.